I've had lots of problems with the last (saved) check number in a register window. When I open a register, tab into the number field, and hit the plus key, invariably the wrong number appears. Once I've entered a check number in a register, the plus key works properly until I close the register and reopen it.
I've tracked the problem down to the gnc_split_register_save_num_cell routine in the file split-register-model-save.c, and have a question about the logic of that routine. This routine saves the last check number only if the 'trans' field of the saved data is not the same as the blank_trans field for that register's particular book. The problem I see is that any time I am entering a new transaction at the bottom of the register, these two fields match and the check number is not saved. If I go change the check number on an old transaction, then the number is remembered as the starting number (- 1) when the register is reopened. It just occurred to me as I was writing this message that the logic of this test might be reversed. That seems too simple a solution, and I feel like I've missed something somewhere. I'd appreciate any pointers. Thanks. David
signature.asc
Description: This is a digitally signed message part
